NetBSD/sys/arch/powerpc/ibm4xx/intr.c
garbled d974db0ada Merge the ppcoea-renovation branch to HEAD.
This branch was a major cleanup and rototill of many of the various OEA
cpu based PPC ports that focused on sharing as much code as possible
between the various ports to eliminate near-identical copies of files in
every tree.  Additionally there is a new PIC system that unifies the
interface to interrupt code for all different OEA ppc arches.  The work
for this branch was done by a variety of people, too long to list here.

TODO:
bebox still needs work to complete the transition to -renovation.
ofppc still needs a bunch of work, which I will be looking at.
ev64260 still needs to be renovated
amigappc was not attempted.

NOTES:
pmppc was removed as an arch, and moved to a evbppc target.
2007-10-17 19:52:51 +00:00

698 lines
16 KiB
C

/* $NetBSD: intr.c,v 1.18 2007/10/17 19:56:39 garbled Exp $ */
/*
* Copyright 2002 Wasabi Systems, Inc.
* All rights reserved.
*
* Written by Eduardo Horvath and Simon Burge for Wasabi Systems, Inc.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed for the NetBSD Project by
* Wasabi Systems, Inc.
* 4. The name of Wasabi Systems, Inc. may not be used to endorse
* or promote products derived from this software without specific prior
* written permission.
*
* THIS SOFTWARE IS PROVIDED BY WASABI SYSTEMS, INC. ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
* TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL WASABI SYSTEMS, INC
* BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: intr.c,v 1.18 2007/10/17 19:56:39 garbled Exp $");
#include <sys/param.h>
#include <sys/malloc.h>
#include <sys/kernel.h>
#include <sys/evcnt.h>
#include <uvm/uvm_extern.h>
#include <machine/intr.h>
#include <machine/psl.h>
#include <powerpc/cpu.h>
#include <powerpc/spr.h>
#include <powerpc/softintr.h>
/*
* Number of interrupts (hard + soft), irq number legality test,
* mapping of irq number to mask and a way to pick irq number
* off a mask of active intrs.
*/
#define ICU_LEN 32
#define LEGAL_IRQ(x) ((x) >= 0 && (x) < ICU_LEN)
#define IRQ_TO_MASK(irq) (0x80000000UL >> (irq))
#define IRQ_OF_MASK(mask) cntlzw(mask)
/*
* Assign these to unused (reserved) interrupt bits.
*
* For 405GP (and 403CGX?) interrupt bits 0-18 and 25-31 are used
* by hardware. This leaves us bits 19-24 for software.
*/
#define IRQ_SOFTNET 19
#define IRQ_SOFTCLOCK 20
#define IRQ_SOFTSERIAL 21
#define IRQ_CLOCK 22
#define IRQ_STATCLOCK 23
/*
* Platform specific code may override any of the above.
*/
#ifdef PPC_IBM403
#include <powerpc/ibm4xx/dcr403cgx.h>
#define INTR_STATUS DCR_EXISR
#define INTR_ACK DCR_EXISR
#define INTR_ENABLE DCR_EXIER
#elif defined(__virtex__)
#include <evbppc/virtex/dev/xintcreg.h>
#define INTR_STATUS XINTC_ISR
#define INTR_ACK XINTC_IAR
#define INTR_ENABLE XINTC_IER
#define INTR_MASTER XINTC_MER
#define INTR_MASTER_EN (MER_HIE|MER_ME)
#undef IRQ_TO_MASK
#undef IRQ_OF_MASK
#undef IRQ_SOFTNET
#undef IRQ_SOFTCLOCK
#undef IRQ_SOFTSERIAL
#undef IRQ_CLOCK
#undef IRQ_STATCLOCK
#define IRQ_TO_MASK(i) (1 << (i)) /* Redefine mappings */
#define IRQ_OF_MASK(m) (31 - cntlzw(m))
#define IRQ_SOFTNET 31 /* Redefine "unused" pins */
#define IRQ_SOFTCLOCK 30
#define IRQ_SOFTSERIAL 29
#define IRQ_CLOCK 28
#define IRQ_STATCLOCK 27
#else /* Generic 405 Universal Interrupt Controller */
#include <powerpc/ibm4xx/dcr405gp.h>
#define INTR_STATUS DCR_UIC0_MSR
#define INTR_ACK DCR_UIC0_SR
#define INTR_ENABLE DCR_UIC0_ER
#endif
#define MASK_SOFTNET IRQ_TO_MASK(IRQ_SOFTNET)
#define MASK_SOFTCLOCK IRQ_TO_MASK(IRQ_SOFTCLOCK)
#define MASK_SOFTSERIAL IRQ_TO_MASK(IRQ_SOFTSERIAL)
#define MASK_STATCLOCK IRQ_TO_MASK(IRQ_STATCLOCK)
#define MASK_CLOCK (IRQ_TO_MASK(IRQ_CLOCK) | IRQ_TO_MASK(IRQ_STATCLOCK))
#define MASK_SOFTINTR (MASK_SOFTCLOCK|MASK_SOFTNET|MASK_SOFTSERIAL)
#define MASK_HARDINTR ~(MASK_SOFTINTR|MASK_CLOCK)
static inline void disable_irq(int);
static inline void enable_irq(int);
static void intr_calculatemasks(void);
static void do_pending_int(void);
static const char *intr_typename(int);
/*
* Interrupt handler chains. intr_establish() inserts a handler into
* the list. The handler is called with its (single) argument.
*/
struct intrhand {
int (*ih_fun)(void *);
void *ih_arg;
struct intrhand *ih_next;
int ih_level;
};
struct intrsrc {
struct evcnt is_evcnt;
struct intrhand *is_head;
u_int is_mask;
int is_level; /* spls bitmask */
int is_type; /* sensitivity */
};
volatile u_int imask[NIPL];
const int mask_clock = MASK_CLOCK;
const int mask_statclock = MASK_STATCLOCK;
static struct intrsrc intrs[ICU_LEN] = {
#define DEFINTR(name) \
{ EVCNT_INITIALIZER(EVCNT_TYPE_INTR, NULL, "uic", name), NULL, 0, 0 }
DEFINTR("pin0"), DEFINTR("pin1"), DEFINTR("pin2"),
DEFINTR("pin3"), DEFINTR("pin4"), DEFINTR("pin5"),
DEFINTR("pin6"), DEFINTR("pin7"), DEFINTR("pin8"),
DEFINTR("pin9"), DEFINTR("pin10"), DEFINTR("pin11"),
DEFINTR("pin12"), DEFINTR("pin13"), DEFINTR("pin14"),
DEFINTR("pin15"), DEFINTR("pin16"), DEFINTR("pin17"),
DEFINTR("pin18"),
/* Reserved intrs, accounted in cpu_info */
DEFINTR(NULL), /* unused "pin19", softnet */
DEFINTR(NULL), /* unused "pin20", softclock */
DEFINTR(NULL), /* unused "pin21", softserial */
DEFINTR(NULL), /* unused "pin22", PIT hardclock */
DEFINTR(NULL), /* unused "pin23", FIT statclock */
DEFINTR("pin24"), DEFINTR("pin25"), DEFINTR("pin26"),
DEFINTR("pin27"), DEFINTR("pin28"), DEFINTR("pin29"),
DEFINTR("pin30"), DEFINTR("pin31")
#undef DEFINTR
};
/* Write External Enable Immediate */
#define wrteei(en) __asm volatile ("wrteei %0" : : "K"(en))
/* Enforce In Order Execution Of I/O */
#define eieio() __asm volatile ("eieio")
/*
* Set up interrupt mapping array.
*/
void
intr_init(void)
{
int i;
for (i = 0; i < ICU_LEN; i++)
switch (i) {
case IRQ_SOFTNET:
case IRQ_SOFTCLOCK:
case IRQ_SOFTSERIAL:
case IRQ_CLOCK:
case IRQ_STATCLOCK:
continue;
default:
evcnt_attach_static(&intrs[i].is_evcnt);
}
/* Initialized in powerpc/ibm4xx/cpu.c */
evcnt_attach_static(&curcpu()->ci_ev_softclock);
evcnt_attach_static(&curcpu()->ci_ev_softnet);
evcnt_attach_static(&curcpu()->ci_ev_softserial);
softintr__init();
mtdcr(INTR_ENABLE, 0x00000000); /* mask all */
mtdcr(INTR_ACK, 0xffffffff); /* acknowledge all */
#ifdef INTR_MASTER
mtdcr(INTR_MASTER, INTR_MASTER_EN); /* enable controller */
#endif
}
/*
* external interrupt handler
*/
void
ext_intr(void)
{
struct cpu_info *ci = curcpu();
struct intrhand *ih;
int i, bits_to_clear;
int r_imen, msr;
int pcpl;
u_long int_state;
pcpl = ci->ci_cpl;
msr = mfmsr();
int_state = mfdcr(INTR_STATUS); /* Read non-masked interrupt status */
bits_to_clear = int_state;
while (int_state) {
i = IRQ_OF_MASK(int_state);
r_imen = IRQ_TO_MASK(i);
int_state &= ~r_imen;
if ((pcpl & r_imen) != 0) {
/* Masked! Mark as pending */
ci->ci_ipending |= r_imen;
disable_irq(i);
} else {
splraise(intrs[i].is_mask);
if (intrs[i].is_type == IST_LEVEL)
disable_irq(i);
wrteei(1);
KERNEL_LOCK(1, NULL);
ih = intrs[i].is_head;
while (ih) {
(*ih->ih_fun)(ih->ih_arg);
ih = ih->ih_next;
}
KERNEL_UNLOCK_ONE(NULL);
mtmsr(msr);
if (intrs[i].is_type == IST_LEVEL)
enable_irq(i);
ci->ci_cpl = pcpl;
uvmexp.intrs++;
intrs[i].is_evcnt.ev_count++;
}
}
mtdcr(INTR_ACK, bits_to_clear); /* Acknowledge all pending interrupts */
wrteei(1);
splx(pcpl);
mtmsr(msr);
}
static inline void
disable_irq(int irq)
{
int mask, omask;
mask = omask = mfdcr(INTR_ENABLE);
mask &= ~IRQ_TO_MASK(irq);
if (mask == omask)
return;
mtdcr(INTR_ENABLE, mask);
#ifdef IRQ_DEBUG
printf("irq_disable: irq=%d, mask=%08x\n",irq,mask);
#endif
}
static inline void
enable_irq(int irq)
{
int mask, omask;
mask = omask = mfdcr(INTR_ENABLE);
mask |= IRQ_TO_MASK(irq);
if (mask == omask)
return;
mtdcr(INTR_ENABLE, mask);
#ifdef IRQ_DEBUG
printf("enable_irq: irq=%d, mask=%08x\n",irq,mask);
#endif
}
static const char *
intr_typename(int type)
{
switch (type) {
case IST_NONE :
return ("none");
case IST_PULSE:
return ("pulsed");
case IST_EDGE:
return ("edge-triggered");
case IST_LEVEL:
return ("level-triggered");
default:
panic("intr_typename: invalid type %d", type);
}
}
/*
* Register an interrupt handler.
*/
void *
intr_establish(int irq, int type, int level, int (*ih_fun)(void *),
void *ih_arg)
{
struct intrhand *ih;
int msr;
if (! LEGAL_IRQ(irq))
panic("intr_establish: bogus irq %d", irq);
if (type == IST_NONE)
panic("intr_establish: bogus type %d for irq %d", type, irq);
/* No point in sleeping unless someone can free memory. */
ih = malloc(sizeof *ih, M_DEVBUF, cold ? M_NOWAIT : M_WAITOK);
if (ih == NULL)
panic("intr_establish: can't malloc handler info");
switch (intrs[irq].is_type) {
case IST_NONE:
intrs[irq].is_type = type;
break;
case IST_EDGE:
case IST_LEVEL:
if (type == intrs[irq].is_type)
break;
/* FALLTHROUGH */
case IST_PULSE:
if (type != IST_NONE)
panic("intr_establish: can't share %s with %s",
intr_typename(intrs[irq].is_type),
intr_typename(type));
break;
}
/*
* We're not on critical paths, so just block intrs for a while.
* Note that spl*() at this point would use old (wrong) masks.
*/
msr = mfmsr();
wrteei(0);
/*
* Poke the real handler in now. We deliberately don't preserve order,
* the user is not allowed to make any assumptions about it anyway.
*/
ih->ih_fun = ih_fun;
ih->ih_arg = ih_arg;
ih->ih_level = level;
ih->ih_next = intrs[irq].is_head;
intrs[irq].is_head = ih;
intr_calculatemasks();
eieio();
mtmsr(msr);
#ifdef IRQ_DEBUG
printf("***** intr_establish: irq%d h=%p arg=%p\n",irq, ih_fun, ih_arg);
#endif
return (ih);
}
/*
* Deregister an interrupt handler.
*/
void
intr_disestablish(void *arg)
{
struct intrhand *ih = arg;
struct intrhand **p;
int i, msr;
/* Lookup the handler. This is expensive, but not run often. */
for (i = 0; i < ICU_LEN; i++)
for (p = &intrs[i].is_head; *p != NULL; p = &(*p)->ih_next)
if (*p == ih)
goto out;
out:
if (i == ICU_LEN)
panic("intr_disestablish: handler not registered");
*p = ih->ih_next;
free(ih, M_DEVBUF);
msr = mfmsr();
wrteei(0);
intr_calculatemasks();
mtmsr(msr);
if (intrs[i].is_head == NULL)
intrs[i].is_type = IST_NONE;
}
/*
* Recalculate the interrupt masks from scratch.
* We could code special registry and deregistry versions of this function that
* would be faster, but the code would be nastier, and we don't expect this to
* happen very much anyway. We assume PSL_EE is clear when we're called.
*/
static void
intr_calculatemasks(void)
{
struct intrhand *q;
int irq, level;
/* First, figure out which levels each IRQ uses. */
for (irq = 0; irq < ICU_LEN; irq++) {
register int levels = 0;
for (q = intrs[irq].is_head; q; q = q->ih_next)
levels |= 1 << q->ih_level;
intrs[irq].is_level = levels;
}
/* Then figure out which IRQs use each level. */
for (level = 0; level < NIPL; level++) {
register int irqs = 0;
for (irq = 0; irq < ICU_LEN; irq++)
if (intrs[irq].is_level & (1 << level))
irqs |= IRQ_TO_MASK(irq);
imask[level] = irqs | MASK_SOFTINTR;
}
/*
* Not external interrupts, make them block themselves manually.
*/
imask[IPL_CLOCK] = MASK_CLOCK;
imask[IPL_STATCLOCK] = MASK_STATCLOCK;
/*
* Initialize the soft interrupt masks to block themselves.
*/
imask[IPL_SOFTCLOCK] = MASK_SOFTCLOCK;
imask[IPL_SOFTNET] = MASK_SOFTNET;
imask[IPL_SOFTSERIAL] = MASK_SOFTSERIAL;
/*
* Enforce hierarchy required by spl(9).
*/
imask[IPL_SOFTNET] |= imask[IPL_SOFTCLOCK];
imask[IPL_SOFTSERIAL] |= imask[IPL_SOFTCLOCK];
/*
* IPL_NONE is used for hardware interrupts that are never blocked,
* and do not block anything else.
*/
imask[IPL_NONE] = 0;
/*
* Enforce a hierarchy that gives slow devices a better chance at not
* dropping data.
*/
imask[IPL_BIO] |= imask[IPL_SOFTNET];
imask[IPL_NET] |= imask[IPL_BIO];
imask[IPL_SOFTSERIAL] |= imask[IPL_NET];
imask[IPL_TTY] |= imask[IPL_SOFTSERIAL];
/*
* There are tty, network and disk drivers that use free() at interrupt
* time, so imp > (tty | net | bio).
*/
imask[IPL_VM] |= imask[IPL_TTY];
imask[IPL_AUDIO] |= imask[IPL_VM];
/*
* Since run queues may be manipulated by both the statclock and tty,
* network, and disk drivers, clock > imp.
*/
imask[IPL_CLOCK] |= imask[IPL_AUDIO];
/*
* We have separate statclock.
*/
imask[IPL_STATCLOCK] |= imask[IPL_CLOCK];
/*
* IPL_HIGH must block everything that can manipulate a run queue.
*/
imask[IPL_HIGH] |= imask[IPL_STATCLOCK];
/*
* We need serial drivers to run at the absolute highest priority to
* avoid overruns, so serial > high.
*/
imask[IPL_SERIAL] |= imask[IPL_HIGH];
/* And eventually calculate the complete masks. */
for (irq = 0; irq < ICU_LEN; irq++) {
register int irqs = IRQ_TO_MASK(irq);
for (q = intrs[irq].is_head; q; q = q->ih_next)
irqs |= imask[q->ih_level];
intrs[irq].is_mask = irqs;
}
for (irq = 0; irq < ICU_LEN; irq++)
if (intrs[irq].is_head != NULL)
enable_irq(irq);
else
disable_irq(irq);
}
static void
do_pending_int(void)
{
struct cpu_info *ci = curcpu();
struct intrhand *ih;
int irq;
int pcpl;
int hwpend;
int emsr;
if (ci->ci_iactive)
return;
ci->ci_iactive = 1;
emsr = mfmsr();
wrteei(0);
pcpl = ci->ci_cpl; /* Turn off all */
again:
while ((hwpend = ci->ci_ipending & ~pcpl & MASK_HARDINTR) != 0) {
irq = IRQ_OF_MASK(hwpend);
if (intrs[irq].is_type != IST_LEVEL)
enable_irq(irq);
ci->ci_ipending &= ~IRQ_TO_MASK(irq);
splraise(intrs[irq].is_mask);
mtmsr(emsr);
KERNEL_LOCK(1, NULL);
ih = intrs[irq].is_head;
while(ih) {
(*ih->ih_fun)(ih->ih_arg);
ih = ih->ih_next;
}
KERNEL_UNLOCK_ONE(NULL);
wrteei(0);
if (intrs[irq].is_type == IST_LEVEL)
enable_irq(irq);
ci->ci_cpl = pcpl;
intrs[irq].is_evcnt.ev_count++;
}
if ((ci->ci_ipending & ~pcpl) & MASK_SOFTSERIAL) {
ci->ci_ipending &= ~MASK_SOFTSERIAL;
splsoftserial();
mtmsr(emsr);
KERNEL_LOCK(1, NULL);
softintr__run(IPL_SOFTSERIAL);
KERNEL_UNLOCK_ONE(NULL);
wrteei(0);
ci->ci_cpl = pcpl;
ci->ci_ev_softserial.ev_count++;
goto again;
}
if ((ci->ci_ipending & ~pcpl) & MASK_SOFTNET) {
ci->ci_ipending &= ~MASK_SOFTNET;
splsoftnet();
mtmsr(emsr);
KERNEL_LOCK(1, NULL);
softintr__run(IPL_SOFTNET);
KERNEL_UNLOCK_ONE(NULL);
wrteei(0);
ci->ci_cpl = pcpl;
ci->ci_ev_softnet.ev_count++;
goto again;
}
if ((ci->ci_ipending & ~pcpl) & MASK_SOFTCLOCK) {
ci->ci_ipending &= ~MASK_SOFTCLOCK;
splsoftclock();
mtmsr(emsr);
KERNEL_LOCK(1, NULL);
softintr__run(IPL_SOFTCLOCK);
KERNEL_UNLOCK_ONE(NULL);
wrteei(0);
ci->ci_cpl = pcpl;
ci->ci_ev_softclock.ev_count++;
goto again;
}
ci->ci_cpl = pcpl; /* Don't use splx... we are here already! */
mtmsr(emsr);
ci->ci_iactive = 0;
}
void
softintr(int idx)
{
static const int softmap[3] = {
MASK_SOFTCLOCK, MASK_SOFTNET, MASK_SOFTSERIAL
};
int oldmsr;
KASSERT(idx >= 0 && idx < 3);
/*
* This could be implemented with lwarx/stwcx to avoid the
* disable/enable...
*/
oldmsr = mfmsr();
wrteei(0);
curcpu()->ci_ipending |= softmap[idx];
mtmsr(oldmsr);
}
int
splraise(int newcpl)
{
struct cpu_info *ci = curcpu();
int oldcpl, oldmsr;
/*
* We're about to block some intrs, so make sure they don't
* fire while we're busy.
*/
oldmsr = mfmsr();
wrteei(0);
oldcpl = ci->ci_cpl;
ci->ci_cpl |= newcpl;
mtmsr(oldmsr);
return (oldcpl);
}
void
splx(int newcpl)
{
struct cpu_info *ci = curcpu();
ci->ci_cpl = newcpl;
if (ci->ci_ipending & ~newcpl)
do_pending_int();
}
int
spllower(int newcpl)
{
struct cpu_info *ci = curcpu();
int oldcpl;
oldcpl = ci->ci_cpl;
ci->ci_cpl = newcpl;
if (ci->ci_ipending & ~newcpl)
do_pending_int();
return (oldcpl);
}